Axillary sampling (AS) and sentinel node biopsy are now accepted standards women with clinically node negative (cN0) breast cancer.
Intraoperative evaluation of the nodes with frozen section (FS) is practiced in many centers.
The procedure is associated with minimal morbidity, is cost effective and obviates the need for re-operation as an axillary clearance can be performed under the same anesthesia after a positive intraoperative diagnosis. Also in negative cases, axillary lymph node dissection and its associated morbidity can beavoided. However, this procedure can be practiced if the false negative or negative predictive value of FS is within acceptable limits. We performed a retrospective audit to evaluate the false negative rate of FS.
We performed a retrospective audit of 239 women with cN0 axilla who underwent axillary sampling with frozen section at our center from May 2014 to April 2015. We compared the frozen section evaluation with the final histopathology (HP) of the nodes dissected in sampling. Additionally, if axillary clearance was done the final nodal positivity was assessed on HP.
A total of 239 cN0 women underwent AS with FS from May 2014 to April 2015 at our institution. The mean pT size was 2.88cm (median 2.5 cm), median age was 51 yrs years and there was 28.03% nodal positivity (67/239) noted on HP. Of these, 182/239 women had intraoperative frozen section diagnosis of negative nodes. Of the FS negative, 6/182 AS were
positive on final HP (falsely labeled negative on FS) and an additional 5/182
had nodes positive in the rest of the axilla (false negative AS). The 6 women
with false negative FS, had a median 8.5 nodes (mean 8.16) dissected in HP (range 5-10) with a median 1.5 (mean1.66) nodes positive (range 1-3) on HP. The sensitivity of FS is 90.32%, negative predictive value (NPV) of 96.72 % and false negative rate (FNR) of 9.67%.
Intraoperative frozen section has a NPV of 96.72% with an FNR of 9.67% and can be a reliable and accurate technique in the hands of an experienced pathologist.
Clinical trial identification
All authors have declared no conflicts of interest.